Environment: - Windows Server 2003 R2 - VSE 8.5, patch 8 - File server w/ roughly 2 TB of data, 500 users. Data replicated to remote site via DSF.
We have been experiencing extreme performance issues on a file server recently, specifically with users copying data down from, or up to this particular fileserver. In some cases we'll see a 5MB file take 90+ seconds, yet at other times it takes 2 or 3 seconds.
After working with Microsoft for a number of days, they are pointing the finger back at McAfee. Our McAfee policies for servers are in-line with best practices (on access on write, dont scan files marked for backup, etc). MS believes the VSE filter driver is causing some latency during file i/o operations and have suggested that the VSE product be removed (...really?).
I am hoping anyone would be willing to post any tweaks or other similar fixes they might have seen in similiar situations with heavily utilized file servers...
Hey BXS, we're experiencing similiar issues, usually start of our production day. Do you have any EVENT ID 1008 and 5051's appearing in your sys/app logs? We noticed that McAfee is terminating itself after a timeout.
We're on 8.5i, but only hotfix 6, so we're going to update to the latest hotfixes and see if this resolves our issue. 2k3 Server, 2tb hard drive, 200 users (less actually at the time of the errors)
5051 txt starts off like this.
A thread in process C:\Program Files\McAfee\VirusScan Enterprise\Mcshield.exe took longer than 90000 ms to complete a request. The process will be terminated. Thread id : 1680 (0x690) Thread address : 0x7C8285EC Thread message :
Some more news, we updated from patch 6.1 to 8 last friday. Monday and tuesday were fine, no issues until today. I noticed that the agent logged a new dat file update during this time. After the update it kicked off a scan, about two minutes later we received another crash of the McShield service and about 15 minutes of connection loss to the server. You could ping the server however we were unable to RDP into it.